Hydrographic data (CTD and discrete bottle data: oxygen, salinity and nutrients) from selected non-WOCE sections and publically available WOCE Hydrographic Program data were gridded using a modified version of Roemmich's (1983) application of objective mapping to hydrographic station data. The plots and gridded data files are a uniform set for the Pacific Ocean only. For the Atlantic and Indian, plots were made for my own use over about one year and so they are not computed uniformly; gridded data files are not available for all and so are not included yet. Prior to the end of 1996 all will be recomputed and placed online.
Downloading the postscript files: if your web browser attempts to plot the postscript files and you want to save them instead, instructions for netscape are: click with the 3rd right mouse button instead of the leftmost one and answer the dialog question.
Bottle data for each of the Atlantic and Indian sections is made available through this site, in NODC's format and in a straight ascii table which does not include all of the header information in the NODC files but which is easier to manipulate.
The public WHP data sets are available through the WHP Special Analysis Centre in Hamburg and at the WHP SAC mirror site at Scripps Institution of Oceanography.
The non-WHP CTD data sets are available in the program "ctdsearch" on the nemo data server at SIO through a public login with a menu.
It is assumed that the chief scientist chose appropriate horizontal sampling,
with tighter spacing across regions of strong gradients.
CTD and discrete data are mapped differently to account for the sampling
differences. The same decay scale and horizontal/vertical scale ratio
are used for both.
CTD data: The 2 dbar profiles are first smoothed vertically with a
Gaussian with half-width 10 dbar. Data are used from 6 stations at a
time, and 6 depths. The mean which is removed is just that of this
small rectangle.
Discrete bottle data: Data from 6 stations are used at one time. An
Akima cubic spline is fit to the deepest station of the group, and this
profile is removed from all stations in the group as a representation of
the mean for the group. The residuals are then mapped and the mean
added back in.
The programs for gridding the data are available from
Lynne Talley (ltalley@ucsd.edu).
Methods
Gridding
The original station data were either CTD profiles, mostly at 2 dbar
intervals, or discrete bottle data, with up to 36 bottles per station.
The data were objectively mapped using Roemmich's (1983) method adapted
for producing sections which subjectively were most like those which are
hand contoured, and hence smoothed somewhat by eye.
Potential density calculations
The potential densities calculated
relative to 0, 2000 and 4000 dbar use potential temperature referenced
to these pressures and then density referenced to these pressures,
respectively. The McDougall neutral density parameter is also used.
References
Jackett, D. R. and T. J. McDougall, 1996. A
neutral density variabile for the world's oceans. J. Phys. Oceanogr., submitted.
Roemmich, D., 1983. Optimal estimation of hydrographic
station data and derived fields. J. Phys. Oceanogr., 13, 1544-1549.
Back to table of contents
Lynne D. Talley, ltalley@ucsd.edu